March, 1946
n-ALKYLTRIMETHYLAND n-ALKYLTRIETHYL-SILANES
tures employed in plasma fractionation have not exceeded mole fraction 0.163 and rarely 0,091, and the temperature has been maintained between 0' and the freezing point of the solution. The balance between (1) the precipitating action of ethanol and ( 2 ) the interaction with salt permits attainment of a variety of conditions under which the protein to be separated may be brought to any desired solubility. This balance is different a t constant ethanol and salt concentration for each pH and temperature, and for each protein component. The pH has been controlled by buffering the ethanol-water mixtures with acetates or other buffers of known ionic strength, and adjusted so as to take advantage of the differences in the iso-
[CONTRIBUTION
FROM THE
475
electric points and the directions of the interactions with salts of the protein components to be separated. Crystallization of a separated component has also been effected in an ethanol-water mixture of low ionic strength and low temperature. Removal of salts and other organic molecules by trituration, reprecipitation, or recrystallization in an ethanol-water mixture a t low temperature replaces dialysis of protein solutions in this system. Removal of the organic precipitant without raising the temperature has generally been accomplished by drying the separated proteins from the frozen state yielding stable, salt-poor, dried protein products of any desired degree of purity. BOSTON,MASSACHUSETTS RECEIVED DECEMBER 12, 1945
SCHOOL OF CHEMISTRY ASD PHYSICS OF THEPENNSYLVANIA STATECOLLEGE
Organo-silicon Compounds. I. Synthesis and Properties of n-Alkyltrimethyl- and n-Alkyltriethyl-silanes DY F. C. WHITMORE, L. H. SOMMER, P. A. DI GIORGIO, IV, A. STRONG, R. E. VANSTRIEN, D. L. BAILEY, H. K. HALL,E. W. PIETRUSZA AND G. T. KERR In 1863 Friedel and Crafts' synthesized tetraethylsilane, the first tetraalkylsilane, by heating diethyl zinc with silicon tetrachloride in a sealed tube; tetramethylsilane was synthesized by the same method2 Twenty years later Polis3 utilized the reaction of sodium with a mixture of aryl halide and silicon tetrachloride for the synthesis of arylsilicon compounds, a method applied later by Kipping4 t o the synthesis of alkylsilicon compounds. In 1904 Kipping5 reported that Grignard reagents could be used in the synthesis of organo-silicon compounds. Chemists were quick to adopt this method, since it was far superior to those previously used, and the synthesis of organosilicon conipounds went forward at a rapid pace. In 1916 Bygden6 reported the preparation and physical properties of sixteen tetraalkylsilanes. An;ong these were five n-alkyltriniethylsilanes and threc n-alkyltriethylsilaries. As a starting point for our program of research on organo-silicon coinpounds, we chose to extend the series of n-alkyltrirnethylsilaxies and n-alkyltriethylsilaries into a higher molecular weight range; a study of the physical properties of such coinpounds seeined of considerable interest. It should be noted that only six carbon analogs of the iiineteen tetraalkylsilanes herein reported are known. The difficulties encountered in the syn(1) Friedel a n d Crafts, Ann., 127, 31 (1863). 12) Friedel and Crafts.. ibid... 136.. 203 (1865). , . i 3 j Polis, Bey., 18, 1540 (1885). (4) Kipping and Lloyd, J. Chem. SOL.,79, 449 (1901). (5) Kipping, Proc. Chem. Soc., 2 0 , 15 (1904). (6) See Krause a n d Grosse "Die Chemie der metall-organischen Verbindungen," GebrQder Borntraeger, Berlin, 1937, p. 263.
thesis of neo hydrocarbons are well known? whereas the synthesis of tetraalkylsilanes, which contain a neo silicon atom (silicon attached to four carbons), is a relatively simple procedure and may be achieved in high yield, due to the ease with which silicon-halogen bonds couple with primary Grignard reagents. The higher n-alkyltrimethyland rt-alkyltriethyl-silanes may possibly supply indications regarding the physical properties of neo hydrocarbons which have yet to be prepared. We have synthesized and here reported seven new n-alkyltrimethylsilanes and six new n-alkyltriethylsilanes, and, in addition, we have prepared again three known n-alkyltrimethylsilanes and three known n-alkyltriethylsilanes to insure a comparable purity for determination of physical properties including viscosities. Incidental to the synthesis of new n-alkyltrimethylsilanes, we have synthesized five new n-alkyltrichlorosilanes. Preparation of the Tetraalkylsilanes In the preparation of n-alkyltrimethylsilanes two methods were used. Method I involved preparation of an n-alkyltrichlorosilane from a Grignard reagent a i d silicon tetrachloride in yields of about SOYo. The n-alkyltrichlorosilane, after purification by fractional distillation, was added to an ether solution of methylmagnesium bromide to give the n-alkyltrimethylsilane in about 70% yield. This was the method used by Bygden6 for the preparation of n-alky-ltrimethylsilanes
+
RMgBr Sic14 +RSiCl, RSiC13 3CHIMgBr --f RSi(CH3)3
+
(7) Cf.Whitmore and Fleming, THISJOURNAL, 65, 4161 (1933).
476 WHITMORE, SOMMER, DI GIORGIO,STRONG, VANSTRIEN, BAILEY,HALL,PIETRUSZA, KERRVol. 68 TABLEI PHYSICAL PROPERTIES OF THE TETRAALKYLSILANES Calcd. molecuNO.
1 2 3 4
Name Ethyl! riniethylsilane
Fropyltrimethylsilane Rutyl1.rinethylsilane Amyltriniethylsilane J Hexyltrimethylsilane 6 Heptyltr: methylsilane 7 Clctyltrimethylsilane 8 Decyltrirnethylsilane 9 Lauryltrimethylsilane 10 31yristylrrimethylsilane 11 ~.lethyltriethglsilane 12 Tetraethgisilane 1 3 Propyltriethylsilane 14 Rutyitriethylsilane 15 Amyltrie :hysilane 16 Hesyltrie thylsilane 17 IIeptyltriethylsilane 1s Octyltrie :hylsilane 10 I)ecpllriethylsilane
lar weight 102.2 116 2 130 3 144.3 158.3 172.4 186.4 214.4 242.4
270.5 130,3 144.3 158.3 172.4 186.4 200.4 214 4 228.4
256.3
Boiling point,
'C.
200 mm. 25 49 74 95 117 137 156 191 222 250 84 108 127 144 162 179 19ei 20s 24 1
760 mm. 62 90 113 139 163 184 202 240 273 300 127 153 173 192 211 230 247 20-2 203
Calcd. heat of
vapn., cal./ mole 7,200 7,500 8,000 9,200 9,800 10,GOO 11,700 12,800 14,100 15,600 8,600 9,500 9,400 11,000 ii,m 11,700 11,700 i 2 , m 14,800
Ref. index 11%
1.3820 1.3929 1.4030 1.4096 1.4154 1 4201 1,4242 1.4310 1.4358 1,4410 1.4160 1.4268 1,4308 1 434s 1.4377 1 .u n o 1.4422 1.4438
1,4472
Method 11, which was used for the synthesis of lauryltriniethylsilane and heptyltrimethylsilane, involved the reaction of trimethylchlorosilane with the appropriate Grignard reagent (CH,),SiCl
-+ RMgBr --+
RSi(CH,i,
In our opinion this latter method of synthesis is better for the following reasons: (a) It is, in essence, a one-step process, since a stock of trimethylchlorosilane can be prepared and used as starting rr aterial for the synthesis of a large number of 72-alkyltrimethylsilanes. (b) It avoids the preparation and purification of high-boiling alkyltrichlorosilanes which are unpleasant to work with due to their ease of hydrolysis t o solid siliconic acids, and (c) it is more economical of the expensive long chain alkyl bromides. Estciisi.\-c use of Method I1 was not possible a t the l q i i i n i i i g of this work because we did not have a. good laboratory method for the preparation of large amoiirits of trimethylchlnrosilane. Later, we prepared pure trirnethylchlorosilane in good yield by the reaction of ammonium chloride with a concentrated sulfuric acid solution of hexameth:,ddisiloxane; details will be published shortly. The u-alkyltriethylsilanes have also been prepared by two methods. Method I, which was only used for the synthesis of three out of the eight compounds prepared, involved the reaction of a i l u-a1l;yltrichlorosilaiie with escess ethylrnagnesiur,i bromide. Method I1 utilized the reaction of :I Grignarcl reagent with triethylchlorosilane as follows RhlgBr
+ (C2H:)SSiCl +RSi(C2Hs)3
Triethylch.lorosilane was prepared in high purity (8) (a) VnFublished work of Pray, Sommer, Goldberg and Whitmore; (b) cf. Flood, THISJ O U R N A L , 55, 1735 (1933); (c) the preparation of trimethylchlorosilane by other methods has been reported. See Taylor a n d Walden, i b i d . , 66, S-12 (1914); Gilliam and Sauer. ihid.. 66. I i S B (1944).
-Density dQ
0.7040 ,7197 ,7352 ,7477 .7678 ,7659 .7729 .7848 ,7938 Solid ,7600 ,7810 ,7868
,7931 ,7977 ,8018 .so45 ,8108 ,8175
in g./ml.dlQ d" 0.6849 ,7020 0.6653 .7181 .I3835 ,7313 ,6979 ,7422 ,7100 ,7506 ,7196 ,7381 ,7277 ,7705 ,7413 .7800 ,7515 ,7911 ,7634 ,7437 ,7107 ,7662 ,7322 ,7724 ,7423 ,7786 ,7489 ,7835 ,7545 ,7880 ,7595 ,7907 .7m7 ,7971 .76!16 .803G ,7771
Mol.
refr. MD
34.73 39.49 44.2s 48.85 53.46 5 s . 14 62.77 72.04 81.22 90.30 43.9s 48.34 Ij3.03 z 7 . 7s 62.43 67,06 71.79 76.09 8j.31
Atomic ref. of silicon SiRD 7.23 7.37 7.54 7.49 7.4s
-
i.54 I .ad
_-
7.5s 7.52 7.36 7.24 6.98 7.05
7.18 7.21 7.22 7.23 7.01 6.99
Absolute viscosity, centipoises O'C. 20°C. 6 O O C . 0.422 0 , 3 3 4 ,479 ,380 GjS ,504 870 ,644 1.186 0 . 8 4 7 0.802 1 618 1.105 0.621 2.149 1.412 0 . 7 8 8 3.718 2.261 1.096 6 190 3 . 4 6 9 1.535 Solid 5.108 2.083 0.679 0 , 5 2 4 .855 ,649 ,998 ,738 1.331 0 9.2 1 . 7 4 0 1.185 0 . 6 6 7 2.314 1.515 0 . 8 1 3 3 O , j t 1 . 9 1 8 0.980 8 86.5 2.3.43 1.144 6 . 5 4 1 3 . 7 0 3 1.648
and good yield from ethyl orthosilicate through the disil~xane.~Here, as in the case of the 1%alkyltrimethylsilanes, Method I1 is the better one. All tetraalkylsilanes reported were purified by careful fractionation in columns of 80 or more theoretical plates. Only constant boiling material having a refractive index range of 0.0002 units or less was used for determination of physical properties. Physical Constants In Table I are listed the nineteen tetraalkylsilanes prepared and some of their physical properties. Boiling points a t 760 mm. are extrapolated from vapor-pressure curves drawn from six or more points covering the range 20-735 mm.; 200 mm. values are taken from the curves. Heats of vaporization were calculated by use of the ClausiusClapeyron equation for the range 200 to 730 mm. Refractive indices are accurate to 0 0002 unit. Densities were determined by the use of pycnometers of about 5-cc. capacity and are corrected to the vacuum values. Molecular refraction was calculated by use of the Lorentz-Lorenz equation Atomic refraction of silicon was calculated by diiference from the molecular refraction, using 2.42 for carbon and 1.10 for hydrogen. l'iscosities were determined by the use of Caniion-Fenske viscometers with a precision of =to 2';O. Discussion of the Physical Properties Boiling Point.-The change in boilitig point with molecular weight is about the same for the trimethyl and triethyl compounds as for hydrocarbons of similar molecular-weight range. The increase in boiling point per methylene group for the silicon compounds varies from 28 to 10' on ascending the series; normal alkanes in the same molecular-weight range have a variation of 27 (9) Unpubl~shedwork of DI Giorgio, Strong, Sommer and \\'hit morr
March, 1946
1~-ALKYLTRIMFJTHYLAND 1~-ALKYLTRIETHYL-SILANES
477
attributed to the fact that their molecular weights are greater b y 16 units. The data of Table I1 may also be used to compare the boiling points of tetraalkylsilanes with those of carbon compounds having about the same molecular weight and molecular symmetry. For this purpose, compare silicon compounds 1, 2, 3, 4, 11 and 12 with carbon compounds 2, 3, 4, 5, 12 and 13, respectively, i. e., EtSiMes with PrCMea. Such a comparison shows that the silicon compounds have boiling points which are from 17-18’ lower, despite the fact that their molecular weights are 2 units higher than the carbon compounds. This is probably caused by the greater compactness of the tetraalkylsilane molecules. Density and Refractive Index.-The increase of density and refractive index with molecular weight is about the same for the tetraalkylsilanes as for hydrocarbons of similar molecular-weight range. The increase in density per methylene
Ei E
C
r. c
-
c
0.8000 J 9 13 17 Carbon atoms per moleculc. Fig. 1.-Change of boiling point with molecular weight: 0 ,alkyltrimethylsilanes; 0, alkyltriethylsilanes; ---, n-alkanes.
to 10’. Figure 1 shows the variation of boiling point with inolecular weight for the two series of compounds and for the normal alkanes of about the same molecular weight.1° When trimethyl and triethyl compounds of the same molecular weight are compared, i. e., compounds 3-8 with 11-15 aiid 17 (Table I) i t is found that the trimethyl coinpounds have lower boiling points. Whal carbon aiialogs of trimethyl conipouiids 3-5 are compared with those of triethyl compounds 11-13, it is found that the trimethyl carbon compounds also have lower boiling points. This is shown in Table 11, which compares the boiling points of eight of the tetraalkylsilanes with those of their carbon analogs. The higher boiling points of the silicon compounds may be
Li 8 0.7600 +
.*h c)
0.7200
0.6800
-
-
I/:
I
TABLE I1 BOILINGI’OXKTS DE”SILICON ASD CARBON ANALOGS Tqpe formula
EtTMc:, PrThlrj BuTMea AmTMes (5) HexTMez (11) MeTEts (12) EtTEts (13) PrTEt, (1) (2) (3) (4)
-Boiling points, ‘“2.T = S i T = C
62 90 115 139 163 127 153 173
?50 79 107 130 155 118 139 161
(10) Physical property d a t a for hydrocarbons are taken from Doss, “Physical Constants of the Principal Hydrocarbons,” T h e Texas i‘ompany, 1942
5 9 13 17 Carbon atoms per molecule. Fig. 3.-Change of refractive index with molecular weight: 0 , alkyltrimethylsilaries; 0-0, alkyltrirthylsilanes; 0- - -0, n-alkanes.
478
WHIThTORE, SOMMER, DI GIORGIO, STRONG, VAN STRIEN,
group varies from 0.018 to 0.005 unit on ascending the series of silicon compounds. Refractive index increases from 0.0109 to 0.0025 unit per additional methylene group. Figs. 2 and 3 give the variation of density and refractive index with molecular weight for the tetraalkylsilanes ; included :for comparison are similar plots for the normal a1kanes. When trimethyl and triethyl compounds of the same molecular weight are compared (Table I ) , it is found that the trimethyl con1pound.s have lower densities and refractive indices. ;Similarly, carbon analogs of trimethyl compounds 3-5 have lower densities and refractive indices than carbon analogs of triethyl compounds 11-13. This is shown in Table I11 which compares the densities and refractive indices of eight of the tetraalkylsilanes with those of their carbon analogs. The higher densities and refractive indices of the silicon compounds are doubtless due to their greater molecular weights.
BAILEY,HALL,PIETRUSZA, KERRV O ~68 .
exception of methyltriethylsilane, have higher densities than the carbon compounds. With the exception of methyltriethylsilane, the silicon compounds have refractive indices which are higher or about the same as those of the carbon compounds; the difference in indices for the first two pairs of compounds is too small to be significant. Greater density and refractive index of the silicon compounds compared to carbon compounds of about the same molecular weight may be attributed to the greater compactness of the tetraalkylsilane molecules; in alkanes refractive index and density are generally higher the niore compact the molecule. An interesting point is that the high symmetry of the tetraethylmethane molecule reverses the usual order so that this substance has a higher density and refractive index than methyltriethylsilane. The temperature coefficients of density for the tetraalkylsilanes are about the same as for normal alkanes of corresponding molecular weights. TABLE I11 In view of the variation in the atomic refraction COMPARISON OF DENSITIESAND REFRACTIVE INDICES FOR of silicon, to be discussed below, it was of interest SILICON ASD CARBOS ANALOGS to compare the change of refractive index with --d20---? Y-W~Ddensity in the trimethyl and triethyl compounds Type formula T = Si T = C T = Si T, C with that in the normal alkanes. Figure 4 shows (1) I3t:rhh 0.6849 0.6492 1.3820 1.3688 the similarity in this relationship for the silicon (‘2) I’rTXlea ,7020 ,6739 1.3929 1.3822 and carbon compounds. ( 3 ) mrifea ,7181 ,6953 1.4030 1.3931 Atomic Refraction of Silicon.--As can be seen (4) XmThler ,7313 .7105 1.4096 1.4023 from Table I the atomic refraction values for ( 5 ) I I ~ x T M Q ,7422 .7233 1.4154 1.4084 silicon in the trimethylsilanes, with the exception ,7274 1.4160 1.4078 .7437 (11) LfeTEta of ethyltrimethylsilane, are all higher than for the ,7662 ,7522 1.4268 1.4197 (12) EtTEta triethylsilanes. In the trimethylsilanes the aver.7724 ,7613 1.4308 1.4258 (13) I-’rTEt3 age value for the atomic refraction of silicon is 7.47 Table 111 may be used to compare densities with a mean deviation of 0.09 unit; for the triand refractive indices of tetraalkylsilanes with ethylsilanes the average value is 7.12 with a mean those of ca.rbon compounds having about the same deviation of 0.09 unit. While it may be argued molecular weight and molecular symmetry (cf. that the differences in SiRD among the trimethylabove). This shows that, although their boiling silanes or triethylsilanes themselves are small points are lower, the silicon compounds, with the enough to warrant an assumption of constancy of atomic refraction of silicon-the differences being due to small amounts of impurities in the compounds-the relatively large difference in the 0.8000 average values of SiRD for the two series of compounds can only be attributed to a real variation in the atomic refraction of silicon. The fact that SiRD is different, even in two such similar series of E 0.7600 compounds, is not too surprising. For example, a the atomic refraction of silicon in triniethylchloro3 silane is 6.72; in hexarnethyldisilme it is 5.45. 2 These data indicate the greatcr polarizability or B 0.7200 deformability of the electronic sheaths in the silicon atom as compared to carbon. We are studying the relationship between SiRD and chemical reactivity of various groups attached to silicon. As yet, no definite conclusions are POSI I I I I I sible. 1.3900 1.4100 1.4300 1.4500 Viscosity.-Since we were unable to find any Refractive index, nz0D. viscosity data for silanes, it was of interest to Fig. 4.--Change of refractive index with density: study the viscosities and temperature-viscosity 0 , alkyltrimethylsilanes; +, alkyltriethylsilanes; 0, n- coefficients of the compounds herein reported. alkanes. The increase in viscosity with molecular weight
-
’
Y
.C
n-ALKYLTRIMETHYLAND 12-ALKYLTRIETHYL-SILANES
ililarch, 194(5
479
3.9 P
+ O ' I +0.6
. 3.1 0
c.l U
rd
.P
9 2.3
Lp .L
?
U CI
g 24
1.5
-0.2
0.7
-
I I I I I 5 9 13 17 Carbon atoms per molecule. Fig. 5.-Change of absolute viscosity with molecular weight: 0 , alkyltrimethylsilanes; 0, alkyltriethylsilanes; n-alkanes. I
+,
is about the same for the tetraalkylsilanes as for normal alkanes in the same molecular-weight range; Fig. 5 shows this relationship. As far as could be determined from available viscosity data for alkanes, it was found that on comparing a tetraalkylsilane with all the known alkanes of about the same molecular weight (alkanes having a molecular weight 2 units less than the silicon compound), the viscosity of the tetraalkylsilane, a t a corresponding temperature, is lower; i. e., ethyltrimethylsilane has a lower viscosity than any of the heptanes. Table IV, which compares absolute viscosities of tetraalkylsilanes and alkanes a t 20°, shows that the silicon compounds have viscosities which are about 25% lower than those of the carbon compounds. TABLE IV
ABSOLUTE VISCOSITIES
Si cpd.
EtSihIe, PrSiMea AmSihle; Et8 HexSiMes HexSiEt, OctSiEta
CARBON COMSAME MOLECULAR WEIGHT
O F SILICON AND
POUNDS OF ABOUT THE
C cpd.
PrCMes BuCMes n-CloHz2 n-CloHzz n-C11H24 n-CtdHso n-Ci6Hs4
Viscosity at Z O O , centipoises Si cpd. C cpd.
0.33 .38 .64 .65 .85 1.52 2.34
0.39 .53 .91 .91 1.17 2.18 3.52
The change of viscosity with temperature is about the same for the tetraalkylsilanes as for
1
0.
3.00 3.20 3.40 3.60 Reciprocal abs. temp. X 1000. Fig. 6.-Change of absolute viscosity with temperature: tetraalkylsilanes, 0-0, bottom t o top, compounds 5, 15, 7 , 17, 8, 19; n-alkanes, ---, bottom t o top, undecane, tetradecane, hexadecane.
normal alkanes. Figure 6 shows this for representative tetraalkylsilanes. Previous Work.-To E. A. Bygden6 must go the credit for the first intensive study of the physical properties of tetraalkylsilanes. Among the compounds which he studied were 1, 2, 3, 12, 13 and 14 in Table I. I n general, this data for these compounds and those of the present work are in good agreement. For example, Bygden's average value for SiRD in compounds 1, 2 and 3 is 7.50 with a mean deviation of 0.06 unit; for compounds 12, 13 and 14 his average value is 7.13 with a mean deviation of 0.06 unit. His values for MD agree with ours to within 0.57,. Boiling points agree to within lo. Experimental Starting Materials.-Silicon tetrachloride (Stauffer Chemical Co., Niagara Falls, Pi. Y . ) was used directly without purification ; fractional distillation of a large batch of this material showed it to be better than Qjc;'0 silicon tetrachloride. Methyl and ethyl bromides (Dow Chemical Co., Midland, Mich.) were used without purification. %-Propyland n-butyl bromides were prepared from good grade commercial alcohols by the hydrobromic-sulfuric acid method. %-Amyl, n-hexyl, n-heptyl, n-octyl, lauryl and myristyl bromides were prepared in yields of 70-807, by reaction of the respective alcohols maintained a t 110" with hydrogen bromide gas which was generated from hydrogen and bromine vapor in a quartz tube at 330". All of the bromides were carefully fractionated in columns of 20 or more theoretical plates and only the purest material was used in the preparation of the Grignard reagents. The n-Alkyltrich1orosilaes.-The synthesis of nhexyltrichlorosilane will be described as representative of the method used. A 3-liter three-necked, round-bottomed flask was equipped with a mercury-sealed mechanical stirrer, reflux condenser and dropping funnel. In the
480 WHITMORE, SOMMER, DI GIORGIO,STRONG, VAN STRIEN, BAILEY,HALL,PIETRUSZA, KERRv01.68 flask there was placed 255 g (1.5 moles) of silicon tetrachloride tlis~jolvedin 500 cc. of dry ether and the system was protected from moisture by a concentrated sulfuric acid trap. n-Hexylmagnesium bromide, 1.5 equivalents by titration, in 600 cc. of ether was filtered into the dropping funnel through glass wool and added t o the silicon tetrachloride oker a period of three hours. Addition was accompanied by the formation of magnesium salts which precipitated, and by moderate refluxing of the ether. After four hours of heating on the steam-bath, the reaction product was filtered by suction and the magnesium salts extracted R ith dry ether. Upon removal of the ether by distillation, the crude product was purified by fractional distillation in a 20-plate glass-helix packed column. The trichlorosilanes were analyzed for chlorine content as follows: weighed samples, 0.5-1 g., were dissolved in 25 cc. of mc.thanol and titrated with standard alkali using phenolphthalein; this method is made possible by the easy hydrolysis of silicon-chlorine bonds. Table V gives the pertinent data for the new trichlorosilanes.
TABLE V NEW TRICHLOROSILASES, RSiC13 R-
--B. "C.
p."-
Mm.
Yield,a
70 52
107 120 C&Iii 50 127 98 C6H13 28 46 C8Hl7 119 84 54 CioH2i 183 29 120 3 C12H26 3 48 CiiHa 156 a Boiling points are uncorrected. the quantity of Grignard reagent.
c C h l o r i n e , %Calcd. Found
51.8
50.8 48.2 48.0 43.0 43.1 38.6 38.3 35.1 34.8 31.7 32.1 Yields are based on
Preparation of n-Alkyltrimethylsilanes by Method 1.The synthesis of n-hexyltrimethylsilane will be described as typical. In the usual apparatus (see above) there was prepared rnethylmagnesium bromide, 1.7 equivalents by titration, irc 800 cc. of ether. The flask was surrounded with an ice-bath, and 107 g. (0.5 mole) of n-hexyltrichlorosilane was added during three hours; precipitation of magnesium halide etherate started after one hour. After completion of the addition, the cooling bath was removed and stirring was continued for ten hours a t room temperature. This was followed by refluxing for five hours, after which the ether was distilled and the residue heated on the steam-bath for five hours. The distilled ether was then returned t o the flask, the solid material broken up, and the contents of the flask worked up with water and acid. The ether layer was separated and the water layer extracted with 100 cc. of ether. After distillation of the ether, the residue was washed with two 20-cc. portions of concentrated sulfuric acid in order to remove silicon-oxygen compounds, followed by washing with water and carbonate solution. The crude product was then dried with calcium chloride and fractionated. Method 11.-Trimethylchlorosilane, 40 g. (0.37 mole), was dissolved in 50 cc. of dry ether and added all a t once to a solution 0:: 0.6 equivalent of laurylmagnesium bromide in 300 cc. of ether. From here on the procedure used was the same as above (Method I ) . Preparation of n-Alkyltriethylsi1anes.-The procedures used for the preparation of these compounds by Methods I and I1 were essentially the same as for the trimethyl compounds. However, cooling during the addition of the trichlorosilane t o ethylmagnesium bromide (Method I), and refluxing after addition of triethylchlorosilane t o the Grignard reagent (Method 11) were found to be unnecessary. Analyses.-The tetraalkylsilanes were analyzed for silicon content in the following manner: Small samples, about 0.5-1 g., were fused with 15 g. of sodium peroxide and 1 g. of sucrose in a Parr bomb. The fusion was dissolved in 100 cc. of distilled water, filtered t o remove a small amount of iron hydroxide, and the filtrate evaporated twice with hydrochloric acid t o dehydrate the silica.
Preparations and analyses of the tetraalkylsilanes are summarized in Table VI.
TABLE VI THETETRAALKYLSILASES Cpd. No.
Method of prepn.
Yield,a
% 32
---Silicon, Calcd.
%-
Found
I 27.5 27.3 I 54 24.1 24.1 I 64 21.5 21.3 I 50 19. a 19.6 I 79 17.7 1i.d 6 I1 46 16.3 16.3 I 89 15.1 15.0 8 I 80 13.1 13.0 9 I1 56 11.6 11.2 10 I 50 10.4 10.4 11 I1 60 21.5 21.2 12 b 68 19.5 19.6 13 I 71 17.7 17.8 14 I1 50 16.3 16.1 I1 75 15,l 15.0 15 16 I1 60 . 14.0 14.2 17 I1 68 13.1 12.8 18 I 77 12 3 12.0 I 78 10.9 10.7 19 a Yields are based on the trichlorosilane for Method I , and on the monochlorosilane for Method 11. Prepared from silicon tetrachloride. 1 2 3 4 5
-
Physical Properties.-Boiling points were determined in a modified Cottrell apparatus.11 Determinations were made a t six pressures covering the range 20-735 mm. Temperatures were read on high precision thermometers graduated in 0.2" (Double Diamond, H. B. Instrument Co.) and pressures were corrected. Refractive indices were determined with an Abbe type refractometer. Densities were measured with pycnometers of about 5-cc. capacity. All determinations were checked by running each compound in two different pycnometers. The instruments were calibrated a t the three temperatures using triple-distilled water. All densities are corrected to the vacuum values. Viscosities were determined in Cannon-Fenske viscometers.l2 All determinations were checked by running each compound in two different instruments.
Acknowledgment.-Our thanks are due the Minnesota Mining and Manufacturing Co. and The Miner Laboratories for a research grant which made this work possible. Summary 1. Ten n-alkyltrimethylsilanes and nine nalkyltriethylsilanes have been prepared. The best method of preparation involves reaction of trimethylchlorosilane or triethylchlorosilane with a Grignard reagent. 2 . Physical properties of these compounds have been determined and discussed. 3. Boiling points, refractive indices and densities are higher in the tetraalkylsilanes than in their carbon analogs. This is attributed to the fact that molecular weights in the silicon compounds are greater by sixteen units. (11) Quiggle, Tongberg and Fenske, Ind. Eng. Chem., Anal. E d , 6, 466 (1934). (12) Cannon and Fenske. i b i d . , 10, 297 (1938).
March, 1946
SILICON
ANALOGS OF NEOPENTYL CHLORIDE
4. The silicon compounds have lower viscosities than alkanes of corresponding molecular &.eight. 5. Comparison of trimethyl and triethyl silicon compounds with trimethyl and triethyl alkanes of corresponding molecular weight, i. e., MesSiEt with MesCPr, reveals that the silicon compounds have lower boiling points, higher densities and
[CONTRIBUTION FROM
THE
451
higher refractive indices than the carbon compounds. This is attributed to the greater compactness of the tetraalkylsilanes. 6, The change of viscosity and density with temperature is about the same in tetraalkylsilanes and in alkanes. STATECOLLEGE, PA.
RECEIVED SOVEMBER 5, 1945
SCHOOL OF CHEMISTRY AND PHYSICS OF THEPENNSYLVANIA STATECOLLEGE]
Organo-silicon Compounds. 1I.l Silicon Analogs of Neopentyl Chloride and Neopentyl Iodide. The Alpha Silicon Effectl
c.
BY FRANK WHITMORE AND LEOH. SOhlMER
The rapidly growing importance of organosilicon chemistry suggested a thorough study of haloalkyl silicon compounds as of practical as well as theoretical interest. Only a few of these compounds have been reported.2 Among these are the monochlorinated derivatives of tetraethylsilane.2a The lower boiling of these two isomers was assigned the a-chloroethyltriethylsilane structure. Since we wished first to investigate the properties of a simple a-haloalkyltrialkylsilane, we synthesized Compound I, chloromethyltrimethylsilane (silico-neopentyl ~ h l o r i d e ) . ~
tetramethylsilane goes about as readily as that of n e ~ p e n t a n e . ~It gives Compound I in fair yield. (CH3)dSi
+ Cla +(CH&SiCH?Cl + HCl (1)
Silico-neopentyl chloride (I) is a colorless stable liquid which can be distilled a t atmospheric pressure without decomposition. It readily forms a Grignard reagent, the first aliphatic example of such a compound containing silicon. The synthetic possibilities of this reagent are being developed. CHs Reaction with mercuric chloride gave silicoI neopentylmercuric chloride (chloromercurirnethylCHa-Si-CH2CI (1) trimethylsilane, trimethylsilylrnethylmercuric I chloride). CH3 Silico-neopentyl chloride (I) is less reactive than We are interested in this compound because of the importance of the analogous neopentyl n-hexyl chloride, but far more reactive than the very inert neopentyl chloride. 4a Results with gr~uping.~ We have compared the properties of chloro- nine reaction mixtures justify these important methyltrimethylsilane with those of its carbon conclusions. Unless otherwise indicated each reanalog; neopentyl chloride. Such studies give action was carried out a t the boiling point of the information on the “silicon effect,” the effect of a solvent for four hours. Whereas n-hexyl chloride silicon atom on the reactivity of a functional reacted completely with sodium ethylate in absolute ethanol, (I) reacted only about 40% under group in an attached carbon chain. Tetramethylsilane was prepared from silicon the same conditions. The activity of the two tetrachloride and methylmagnesium b r ~ m i d e . ~ compounds with the following four reagents was The liquid phase, photochemical chlorination of approximately the same. The figures in parentheses indicate the percentage reaction, the first (1) Presented before the Division of Organic Chemistry a t the figure in each case corresponding to chloromethylPittsburgh Meeting of the American Chemical Society, September 6, trimethylsilane (I) and the second to n-hexyl 1043. Paper I, Whitmore, e l al., THISJOURNAL, 68, 475 (1946). chloride : potassium acetate in absolute ethanol (2) (a) Ushakov and Itenberg, J . Gen. Chem. U.S. S . R . , 7, 2r95 (1937); (b) Krieble and Elliot, THISJOURNAL, 67, 1810 (1945). (23-29) ; potassium acetate in glacial acetic acid (3) We use the silico-neopentyl naming as well as that sug(26-32) ; potassium hydroxide in absolute ethanol gested by Sauer, J. Chem. Ed., 21, 303 (1944), and the unpublished (52-69); potassium hydroxide in 70: 30 aqueous Dow Committee Report of July 1 , 1944. ethanol (16-22). It will be noted that (I) ap(4) (a) Whitmore and co-workers, ibid., 6 6 , 4161 (1933); 61, 1585, 1586 (1939); 63, 124 (1941); 64, 1783 (1942); (b) Skell and pears slightly less active in each case. Aqueous Hauser, ibid., 64, 2633 (1942); (c) Aston and co-workers, J . Chem. potassium hydroxide gave no reaction with (I) Phys., 12, 336 (1944); (d) Kharasch and Fineman, THISJOURNAL, but gave about 1% conversion with the alkyl 63, 2776 (1941); (e) van Wijk and others, Physica, 7, 45 (1940); halide. Refluxing for eight hours with aqueous ( f ) Halford, J . Chem. Phys., 8, 496 (1940); (9) Kincaid and Hen62, 1474 (1940); (h) Silver, J . Chem. riques, Jr., THIS JOURNAL, ethanol (1:1) gave no reaction with (I) and less Phys., 7 , 1113 (1939); (i) Frey, C . A , , 37, 5079 (1943). 1% change with the other chloride. Boiling than (5) (a) Krause and von Grosse, “Die Chemie der metall-organi(I) for five minutes with ethanolic silver nitrate schen Verbindungen,” Berlin, 1937. p. 260; (b) Bygden, Bey., 44, gave no action, whereas similar treatment of 2640 (1911).
